This document provides an overview of sentiment analysis and opinion mining. It discusses how sentiment analysis can be used to analyze text and determine the subjective opinions and attitudes expressed. It covers key aspects of sentiment analysis including challenges, common approaches like machine learning classifiers, datasets used, and different levels of analysis like document, sentence, and feature level sentiment classification. The goal of sentiment analysis is to automatically identify positive or negative opinions in text data to understand user sentiment.

1. SENTIMENT ANALYSIS AND
OPINION MINING
Avinash Kumar Singh

2. “WHAT OTHER PEOPLE THINK ?”
 What others think has always been an important
piece of information
 Before making any decision, we look for
suggestions and opinions from others.
 A big question “So whom shall I ask ?”.
2

3. EVOLUTION
History
• Friends
• Acquaintances
• Consumer Reports
Present
• Friends +
Acquaintances
• Unknowns
• No Limitations !
• Across Globe 3

4. MORE PROBLEMS !!
 Biased views
 Fake Reviews
 Spam Reviews
 Contradicting Reviews
4

5. SOLUTION ! – SUBJECTIVITY
ANALYSIS
• General Text can be divided into two
segments
 Objective – which don’t carry any opinion or
sentiment.
 Facts (news, encyclopedias, etc)
 Subjective
• Subjectivity Analysis
 Linguistic expressions of somebody’s opinions,
sentiments, emotions .. that is not open to
verification.
5

6. FLAVORS OF SUBJECTIVITY
ANALYSIS
Sentiment
Analysis
Opinion
Mining
Mood
Classification
Emotion
Analysis
Synonyms and
Used Interchangeably !!
6

7. WHAT IS SENTIMENT?
• Subjective impressions
• Generally, Sentiment ==
 Feelings
 Opinions
 Emotions
 Attitude
• like/dislike or good/bad, etc.
7

8. WHAT IS SENTIMENT ANALYSIS?
• Sentiment Analysis is a study of human
behavior in which we extract user opinion
and emotion from plain text.
• Identifying the orientation of opinions in a
piece of text.
 The Maggie is Tasty [Sentiment] 
 This movie stars Mr. X. [Factual]
 This Maggi is silent killer [Sentiment] 
8

9. MOTIVATION
 Enormous amount of information.
 Real time update
 Monetary benefits
9

10. DOES WEB REALLY CONTAIN
SENTIMENTS ?
 Yes, Where ?
 Blogs
 Reviews
 User Comments
 Discussion Forums
 Social Network (Twitter, Facebook, etc.)
10

11. CHALLENGES
 Ambiguous words
 This music cd is literal waste of time. (negative)
 Please throw your waste material here. (neutral)
 Sarcasm detection and handling
 “All the features you want - too bad they don’t
work. :-P”
 (Almost) No resources and tools for low/scarce
resource languages like Indian languages.
11

12. BASICS ..
 Basic components
 Opinion Holder – Who is talking ?
 Object – Item on which opinion is expressed.
 Opinion – Attitude or view of the opinion holder.
This is a
good book.
Opinion
Holder
Object
Opinion
12

13. TYPES OF OPINIONS
 Direct
 “This is a great book.”
 “Mobile with awesome functions.”
 Comparison
 “Samsung Galaxy S3 is better than Apple iPhone 4S.”
 “Hyundai Eon is not as good as Maruti Alto ! .”
13

14. WHAT IS SENTIMENT
CLASSIFICATION
 Classify given text on the overall sentiments
expresses by the author
 Different levels
 Document
 Sentence
 Feature
 Classification levels
 Binary
 Multi Class
14

15. DOCUMENT LEVEL SENTIMENT
CLASSIFICATION
 Documents can be reviews, blog posts, ..
 Assumption:
 Each document focuses on single object.
 Only single opinion holder.
 Task : determine the overall sentiment
orientation of the document.
15

16. SENTENCE LEVEL SENTIMENT
CLASSIFICATION
 Considers each sentence as a separate unit.
 Assumption : sentence contain only one opinion.
 Task 1: identify if sentence is subjective or
objective
 Task 2: identify polarity of sentence.
16

17. FEATURE LEVEL SENTIMENT
CLASSIFICATION
 Task 1: identify and extract object features
 Task 2: determine polarity of opinions on
features
 Task 3: group same features
 Task 4: summarization
 Ex. This mobile has good camera but poor
battery life.
17

18. APPROACHES
 Prior Learning
 Subjective Lexicon
 (Un)Supervised Machine Learning
18

19. 1.1 KEYWORDS SELECTION FROM
TEXT
 Pang et. al. (2002)
 Two human’s hired to pick keywords
 Binary Classification of Keywords
 Positive
 Negative
 Unigram method reached 80% accuracy.
19

20. 1.2 N-GRAM BASED CLASSIFICATION
 Learn N-Grams (frequencies) from pre-annotated
training data.
 Use this model to classify new incoming sample.
 Classification can be done using
 Counting method
 Scoring function(s)
20

21. 1.3 PART-OF-SPEECH BASED
PATTERNS
 Extract POS patterns from training data.
 Usually used for subjective vs objective
classification.
 Adjectives and Adverbs contain sentiments
 Example patterns
 *-JJ-NN : trigram pattern
 JJ-NNP : bigram pattern
 *-JJ : bigram pattern
21

22. DICTIONARY OF AFFECTIVE
LANGUAGE
 9000 Words with Part-of-speech information
 Each word has a valance score range 1 – 3.
 1 for Negative
 3 for Positive
 App
 http://sail.usc.edu/~kazemzad/emotion_in_text_cgi/DAL_app/inde
22

23. ADVANTAGES AND
DISADVANTAGES
 Advantages
 Fast
 No Training data necessary
 Good initial accuracy
 Disadvantages
 Does not deal with multiple word senses
 Does not work for multiple word phrases
23

24. APPROACH 3: MACHINE LEARNING
 Sensitive to sparse and insufficient data.
 Supervised methods require annotated data.
 Training data is used to create a hyper plane
between the two classes.
 New instances are classified by finding their
position on hyper plane.
24

25. SOME UNANSWERED QUESTIONS !
 Sarcasm Handling
 Word Sense Disambiguation
 Pre-processing and cleaning
 Multi-class classification
25

26. DATASETS
 MPQA Corpus
 Multi Perspective Question Answering
 News Article, other text documents
 Manually annotated
 692 documents
 Twitter Dataset
 http://www.sentiment140.com/
 1.6 million annotated tweets
 Bi-Polar classification
26

27. READING
 Opinion Mining and Sentiment Analysis
 Bo Pang and Lillian Lee (2008)
 www.cs.cornell.edu/home/llee/omsa/omsa.pdf
 Book: Sentiment Analysis and Opinion Mining
 Bing Liu (2012)
 http://www.cs.uic.edu/~liub/FBS/SentimentAnalysis-
and-OpinionMining.html
27

28. THANK YOU